English

Mochi: Aligning Pre-training and Inference for Efficient Graph Foundation Models via Meta-Learning

Machine Learning 2026-05-08 v2 Artificial Intelligence

Abstract

We propose Mochi, a Graph Foundation Model that addresses task unification and training efficiency by adopting a meta-learning based training framework. Prior models pre-train with reconstruction-based objectives such as link prediction, and assume that the resulting representations can be aligned with downstream tasks through a separate unification step such as class prototypes. We demonstrate through synthetic and real-world experiments that this procedure, while simple and intuitive, has limitations that directly affect downstream task performance. To address these limitations, Mochi pre-trains on few-shot episodes that mirror the downstream evaluation protocol, aligning the training objective with inference rather than relying on a post-hoc unification step. We show that Mochi, along with its more powerful variant Mochi++, achieves competitive or superior performance compared to existing Graph Foundation Models across 25 real-world graph datasets spanning node classification, link prediction, and graph classification, while requiring 8\sim27 times less training time than the strongest baseline.

Keywords

Cite

@article{arxiv.2604.22031,
  title  = {Mochi: Aligning Pre-training and Inference for Efficient Graph Foundation Models via Meta-Learning},
  author = {João Mattos and Arlei Silva},
  journal= {arXiv preprint arXiv:2604.22031},
  year   = {2026}
}

Comments

23 pages, 7 figures

R2 v1 2026-07-01T12:33:02.927Z